Role of multiparametric MRI in long-term surveillance following focal laser ablation of prostate cancer

Abstract

Objective:

Determine the multiparametric magnetic resonance imaging (mpMRI) appearance of the prostate following focal laser ablation (FLA) for PCa and to identify imaging characteristics associated with recurrent disease.

Methods:

Retrospective analysis of patients who underwent FLA for low-intermediate risk PCa between 2010 and 2014 was performed. Early (median 4 months) and late mpMRI (median 49 months) follow-up were qualitatively assessed for T ₂-weighted, dynamic contrast enhanced (DCE) and diffusion weighted imaging (DWI) appearances and also compared to corresponding PSA values and biopsy results.

Results:

55 cancers were treated in 54 men (mean age 61.0 years). Early mpMRI was performed in 30 (54.5%) patients while late follow-up mpMRI in 42 (84%). Ill-defined scarring with and without atrophy at the treatment site were the most common appearances. In patients with paired MRI and biopsy, one of four patients with clinically significant PCa on biopsy (≥GG2 or≥6 mm GG1) showed hyperenhancement or restricted diffusion at early follow-up. At late follow-up, positive biopsies were seen in 5/8 (63%) cases with hyperenhancement and 5/6 (83%) cases with restricted diffusion at the treatment site. PSA change was not associated with biopsy results at either time point.

Conclusion:

mpMRI is able to document the morphological and temporal changes following focal therapy. It has limited ability to detect recurrent disease in early months following treatment. Late-term mpMRI is sensitive at identifying patients with recurrent disease. Small sample size is, however, a limitation of the study.

Advances in knowledge:

Implementing MRI in follow-up after FT may be useful in predicting residual or recurrent PCa and therefore provide reliable outcome data.

Introduction

Prostate cancer (PCa) is the second leading cause of cancer-related mortality in males. ¹ Treatment options are varied depending on multiple factors; however, patients are generally stratified into risk groups. For patients with intermediate-high risk PCa whole gland therapy by means of prostatectomy, or radiotherapy is recommended. ² However, conventional procedures carry a high risk of bladder, bowel and sexual dysfunction. ²

For low-risk PCa, the natural history of the disease is less aggressive such that patients may elect to undergo active surveillance. ² While the risk of disease progression is low, metastatic disease and mortality remain possible. ³ Due to these factors and the morbidity associated with radical treatments, new minimally invasive image-guided therapies have been developed in recent years to counter this. ^4,5 Low and intermediate risk PCa localized on mpMRI can be treated with focal therapy (FT). ⁶

Focal laser ablation (FLA) is one such technique which involves the insertion of a laser fibre into the PCa under imaging guidance. Once adequately positioned, diode laser light at a specific wavelength is delivered causing irreversible thermal injury destroying viable tumour in a targeted approach. ^7,8 FLA has been shown to be safe and well tolerated ^5,9–14 although further validation with long-term follow-up is required for widespread adoption. ¹⁵

For patients treated with whole gland therapy, prostate-specific antigen (PSA) monitoring is useful to detect recurrent disease. However, PSA levels are less reliable in patients treated with FT due to a variable reduction in prostate volume post-therapy. ^16,17 Consequently, MRI has been shown to be superior in detecting recurrent disease and its routine use in post-FT surveillance is now endorsed by consensus recommendation groups. ^18,19

The test characteristics of post-FT mpMRI are not fully understood as yet. ²⁰ The main aim of this study was therefore to assess the mpMRI appearance of the prostate following targeted FLA in an attempt to identify specific changes associated with recurrent disease.

Methods and materials

Patient population:

Institutional Review Board ethics approval was obtained for this single-centre retrospective study (REB #18–5046). All patients who underwent FLA for low-intermediate risk localized PCa between January 2010 and September 2014 in institute’s Phase 1 prospective development study were reviewed. Included patients were diagnosed with PCa grade group (GG) ≤ 3 with a visible focal lesion and ≤2 cm on mpMRI.

Inclusion criteria:

Early follow-up was defined as 4–8 months post FT. During this time, mpMRI scans were assessed, PSA values and biopsy results were recorded. Late follow-up was defined according to the most recent mpMRI, or the mpMRI prior to any treatment in patients who underwent further PCa treatment, but >12 months following FT. Only patients with biopsy within 9 months following the mpMRI were included in the study for late analysis Figure 1. PSA values were also recorded.

Figure 1.

Patient Enrolment: mpMRI = multiparametric-parametric magnetic resonance imaging

Treatment technique:

Treatments were performed in-bore using a 1.5T GE Excite Twinspeed MRI (GE Healthcare, Milwaukee, Wisconsin) under deep sedation (intravenous Propofol). In short, 980 nm VisualaseⓇ (Houston, Texas, USA) photothermal laser fibres were inserted transperineally. ²¹ The ablation zone was monitored in real time using custom developed MRI thermography.

Biopsy technique and interpretation:

Early biopsies (3–7 targeted samples from the ablation zone including margins as per study protocol) and late clinical biopsies (12 core systematic biopsy with 3–7 targeted samples of the ablation zone including margins) were performed transrectally under ultrasound guidance, either with cognitive visual or software fusion (Artemis 2.0, Eigen, California). Biopsies prior to 2013 were performed with cognitive assessment of treatment area and identified on real-time TRUS as area of volume loss, scarring and/or retention cyst. For biopsies performed with software fusion, the treatment area and/or the suspicious MRI focus were outlined and marked as region of interest for targeting. Biopsies were considered positive for recurrent disease if ≥6 mm GG1 or any volume ≥GG2 was identified. Biopsies with GG1 (≤5 mm) were deemed to be not clinically significant(cs).

MRI technique:

Patients were scanned using Siemens Avanto 1.5T and Magnetom Verio 3.0T systems (Siemens Medical, Forcheim, Germany). Standard sequences included: T2W axial, coronal and sagittal images; Diffusion-weighted axial images (B-values 0, 100, 400 and 800 s/mm². After 2013, B-values of 1000 s/mm² and interpolated B-value 1400 s/mm² were also obtained; calculated ADC maps; and dynamic contrast-enhanced T1W axial images (3D-FLASH) with temporal resolution of 5.4 s.

MRI analysis:

PCa location and size were recorded. The images were retrospectively evaluated for the following in the treatment area: T2W images were scored as ¹ : Ill-defined scarring, defined as linear hypointense signal ² ; Ill-defined scarring with atrophy, defined as linear hypointense signal with pronounced volume loss ³ ; Well-defined scarring, defined as circumscribed hypointense signal; and ⁴ Scarring with cystic change, defined as ill-defined hypointense signal with areas of hyperintensity. Patients demonstrating both scarring and cystic change with atrophy were ultimately classified as scarring with cystic change. DCE images were scored as ¹ : Hypoenhancement, defined as reduced enhancement of the ablation zone relative to parenchyma ² ; Isoenhancement, defined as homogenous enhancement relative to parenchyma; and ³ Hyperenhancement, defined as focal and early increased enhancement within or along the margins of the ablation zone. DWI was scored as ¹ : no restricted diffusion, defined as anything other than restricted diffusion; and ² Restricted diffusion, defined as hyperintense signal on high B-value images with corresponding ADC hypointensity.

MR images were initially reviewed by a radiologist with 2 years (**) of experience in prostate MRI reporting. Scores were assigned to T2W, DWI and DCE images. All images and scores were reviewed and confirmed with a senior radiologist (**) with >8 years of experience in interpreting prostate MRI for a final consensus read. The readers were blinded to biopsy results.

Statistical analysis:

Descriptive statistics were used to summarize the patient population and imaging parameters. Mean, median and ranges were used to describe quantitative variables, frequencies and proportions for categorical variables. The Fisher’s exact test was used to examine the association between biopsy results and imaging parameters. GEE models were used to compare PSA values between the biopsy results at early and late time periods. P-value of 0.05 was used to determine statistical significance. Statistical analysis was performed using R 3.6.1 (R Core Team v.3.6.1, Vienna, Austria).

Results

Patient characteristics:

In total, 55 cancers in 54 men were treated between January 2010 and September 2014 (mean patient age 61.0 years, range 43.1–77.0 years, mean PSA 5.44, range 0.11–18.31). All lesions were MRI visible. 33 (60%) lesions were GG1 disease, 21 (38%) were GG2 and 1 lesion (2%) was GG3 disease on targeted pretreatment biopsy. The majority of cancers were located within the peripheral zone (51, 93%) with a minority in the transition zone (4, 7%). Mean cancer core length (CCL) on biopsy was 0.85 cm for GG1, 0.89 cm for GG2/3 and 0.87 cm for all cancers. Patient demographics are provided in Table 1

Table 1.

Clinical details for the treated patients

Patient and Tumour Characteristics
Participants	54
Age in years (mean, range)	61 (43.1–77)
Total number of PCa treatment sites	55
Gleason Score at baseline	GG1 – 33 (31 in Pz, two in Tz) GG2 – 21 (19 in Pz, two in Tz) GG3 – 1 (Pz)
Tumour length at baseline in cm (mean, range)	GG1 – 0.85 (0.5–1.8) GG2/3 – 0.89 (0.5–1.9)
PSA ng/ml (mean, range)	5.44 (0.11–18.31)
Prostate volume cc (mean, range)	40 (14.4–268)
PSA density (mean, range), all males GG1 disease (n = 33) GG2/3 disease (n = 22)	0.13 (0.01–0.49) 0.13 (0.01–0.49) 0.13 (0.03–0.29)
PIRADS Score 3 4 5	3 (3 GG1) 47 (27 GG1, 19GG2, 1GG3) 5 (3 GG1, 2 GG2)

Early follow-up:

Paired early mpMRI and biopsy follow-up were performed in 28 cases (52%) at median 4 months (range 3.5–7.7 months). On T2W images, the two most common appearances were ill-defined scarring (36%) and ill-defined scarring with atrophy (32%) Figure 2. On DCE, the ablation zone was most commonly seen as hypoenhancement (54%) followed by isoenhancement (32%) with hyperenhancement seen in a minority (14%) Figure 3. On DWI, only a minority demonstrating restricted diffusion (11%) Figure 4, Table 2.

Figure 2.

T2W images were scored as ¹ : (1) Ill-defined scarring (A), defined as linear hypointense signal; (2) Ill-defined scarring with atrophy (B), defined as linear hypointense signal with pronounced volume loss; (3) Well-defined scarring (C), defined as circumscribed hypointense signal; and (4) Scarring with cystic change (D), defined as ill-defined hypointense signal with areas of hyperintensity. Patients demonstrating both scarring and cystic change and atrophy were ultimately classified as scarring with cystic change.

Figure 3.

DCE images were scored as: (1) Hypoenhancement (A), defined as reduced enhancement of the ablation zone relative to parenchyma; (2) Isoenhancement (B), defined as homogenous enhancement relative to parenchyma; and (3) Hyperenhancement (C), defined as increased enhancement within or along the margins of the ablation zone.

Figure 4.

DWI was scored as: (1) No restricted diffusion (A,) high b value image and B,) ADC map), defined as anything other than restricted diffusion; and (2) Restricted diffusion (C-D), defined as hyperintense B-value images (C) with corresponding ADC hypointensity (D).

Table 2.

Early and late follow-up mpMRI appearance of the prostate. One patient in the late mpMRI follow-up group was unable to receive contrast accounting for one less number in this group. T2 score: (1) Ill-defined scarring, (2) ill-defined scarring with atrophy, (3) well-defined scarring and (4) scarring with cystic change. DCE score: (1) Hypoenhancement, (2) isoenhancement and (3) hyperenhancement. DWI score: (1) No restricted diffusion and (2) restricted diffusion

Appearance	Early MRI (median 4 months)			Late MRI (median 49 months)
T2 score	Negative	Positive	All	Negative	Positive	All
1	8 (80%)	2 (20%)	10 (100%)	6 (67%)	3 (33%)	9 (100%)
2	8 (89%)	1 (11%)	9 (100%)	4 (57%)	3 (43%)	7 (100%)
3	4 (80%)	1 (20%)	5 (100%)	3 (75%)	1 (25%)	4 (100%)
4	4 (100%)	0 (0%)	4 (100%)	0 (0%)	1 (100%)	1 (100%)
Total	24 (86%)	4 (14%)	28 (100%)	13 (62%)	8 (38%)	21 (100%)
DCE score	Negative	Positive	All	Negative	Positive	All
1	13 (87%)	2 (13%)	15 (100%)	4 (80%)	1 (20%)	5 (100%)
2	8 (89%)	1 (11%)	9 (100%)	6 (86%)	1 (14%)	7 (100%)
3	3 (75%)	1 (25%)	4 (100%)	3 (38%)	5 (62%)	8 (100%)
Total	24 (86%)	4 (14%)	28 (100%)	13 (65%)	7 (35%)	20 (100%)
DWI score	Negative	Positive	All	Negative	Positive	All
1	21 (84%)	4 (16%)	25 (100%)	12 (80%)	3 (20%)	15 (100%)
2	3 (100%)	0 (0%)	3 (100%)	1 (17%)	5 (83%)	6 (100%)
Total	24 (86%)	4 (14%)	28 (100%)	13 (50%)	8 (50%)	21 (100%)

Late follow-up:

After early follow-up, five patients (9%) with positive biopsies progressed to whole gland treatment such that 49 patients (91%) proceeded to late clinical follow-up. Late mpMRI (median 49 months, range 21–101 months) was performed in 42 of these patients of which 21 also had a paired biopsy for inclusion in this study. On T2W images, ill-defined scarring (43%) and ill-defined scarring with atrophy (33%) remained the two most common appearances. One patient refused contrast in the late MRI cohort. On DCE, hypoenhancement was seen in 20% while there was a relative even occurrence of isoenhancement (35%) and hyperenhancement (40%). On DWI, 29% showed restricted diffusion Table 2.

Early follow-up - mpMRI appearance and PSA values versus biopsy results:

Post treatment, early follow-up biopsy (targeted to treatment area only) was performed in 51 cases (93%) and mpMRI (median 4 months, range 3.6–7.7 months) in 30 cases (55%) following treatment. Out the 30 cases who underwent MRI, 28 (57%) also underwent biopsy (median 4.3 months, range 3.7–6.2 months). Three of 28 MRI’s were performed following the biopsy (2.7, 3 and 3 months, respectively), but this was blinded to the readers. No haemorrhage was noted in all three scans. In total, five cases (10%) were positive at biopsy, of which four had a mpMRI.

On DCE, focal hyperenhancement relative to the remaining gland at treatment site was present in four cases, three of which were negative at biopsy. Of the 24 cases that did not show hyperenhancement, three (12%) were positive while 21 (88%) were negative. Out of the four positive biopsies, hyperenhancement was present in one case (p = 0.48) Table 3.

Table 3.

Early follow-up comparison of hyperenhancement and restricted diffusion compared against biopsy results. DCE score: (1) Hypoenhancement, (2) isoenhancement and (3) hyperenhancement. DWI score: (1) No restricted diffusion and (2) restricted diffusion

Early follow-up	Early biopsy (median 4.3 months)
DCE scores	Negative	Positive (≥6 mm GG1)	All	p-value
Non-hyperenhancement (1-2)	21 (88%)	3 (12%)	24 (100%)	0.48
Hyperenhancement (3)	3 (75%)	1 (25%)	4 (100%)
Total	24 (86%)	4 (14%)	28 (100%)
DWI score	Negative	Positive	All	p-value
Non-restricting (1)	21 (84%)	4 (16%)	25 (100%)	1.000
Restricting (2)	3 (100%)	0 (0%)	3 (100%)
Total	24 (86%)	4 (14%)	28 (100%)

GG, ISUP Grade group.

On DWI restricted diffusion was present in three cases, all of which were negative at biopsy. Of the 25 cases that did not show restricted diffusion, four (16%) were positive while 21 (84%) were negative (p = 1) Table 3.

For the 28 patients with both mpMRI and biopsy data, 23 of them had PSA values available. Mean PSA declined from 4.2 ng ml⁻¹ (range 1.1–10.7 ng ml⁻¹) to 2.7 ng ml⁻¹ (range 0.4–5.9 ng ml⁻¹) in the early follow-up period. In males with positive biopsy, the mean PSA declined from 4.85 to 3.05 ng ml⁻¹ compared to mean PSA change from 4.09 to 2.63 ng ml⁻¹ in males without csPCa at early biopsy [p = 0.75] Table 4.

Table 4.

Mean PSA values between patients with negative and positive biopsy results at the early and late follow-up after targeted focal therapy

Early follow-up (median 4 months after treatment)
	Negative N = 21	Positive N = 2	P-value
Mean PSA (ng/ml) at baseline	4.09	4.85	0.75 (GEE model)
Mean PSA (ng/ml) at early follow-up	2.63	3.05
Late follow-up (median 53 months after treatment)
	Negative N = 12	Positive N = 5	P-value
Mean PSA (ng/ml) at 4 months following treatment (‘nadir’)	3.94	2.12	0.87 (GEE model)
Mean PSA (ng/ml) at late follow-up	6.99	4.76

Note: The GEE models considered the specific time for early and late follow-up for each patient

Late follow-up: Biopsy results versus MRI appearance and PSA change:

At late follow-up, post-treatment mpMRI was performed in 42 cases (84%) while paired biopsy was performed in 21 of these cases (50%) at median 49 months (21–101 months; late biopsy median time 53 months, range 25–103). Biopsies were positive in treatment area in eight cases (38%), and negative in 13 cases (62%). In the remaining 28 patients, clinical follow-up (biopsy, PSA or mpMRI) was obtained but without paired mpMRI and biopsy and therefore beyond the scope of inclusion in this manuscript.

Overall, MRI detected csPCa in six of eight patients (75%). None of the patients demonstrated new focus of csPCa on MRI or biopsy remote from the treatment site in the time period.

One of the 21 patients in this cohort refused contrast during mpMRI. On DCE, hyperenhancement at the treatment site was present in eight cases, of which five (63%) had a positive biopsy. Out of the 12 cases that did not show hyperenhancement, two cases (17%) were positive while 10 (83%) were negative for csPCa [p = 0.06] Table 5.

Table 5.

Late follow-up comparison of hyperenhancement and restricted diffusion compared against biopsy results. DCE score: (1) Hypoenhancement, (2) isoenhancement and hyperenhancement. DWI score: (1) No restricted diffusion and (2) restricted diffusion

Late follow-up	Late biopsy (median 53 months)
DCE scores	Negative	Positive	All	p-value
Non-hyperenhancement (1-2)	10 (63%)	2 (29%)	12 (100%)	0.06
Hyperenhancement (3)	3 (33%)	5 (71%)	8 (100%)
Total	13 (65%)	7 (35%)	20 (100%)
DWI score	Negative	Positive	All
Non-restricting (1)	12 (92%)	3 (38%)	15 (100%)	0.01
Restricting (2)	1 (8%)	5 (62%)	6 (100%)
Total	13 (62%)	8 (38%)	21 (100%)

On DWI, restricted diffusion was present in six cases, of which five (83%) had a positive biopsy while one (17%) was negative. Out of the 15 cases that did not demonstrate restricted diffusion, three cases (20%) were positive while 12 (80%) were negative. Out of the eight positive biopsies, restricted diffusion was present in five cases (63%) and absent in three (37%). There was statistically significant association between restricted diffusion and recurrent disease [p = 0.01] Table 5.

Independently, the positive-predictive value (PPV) and negative-predictive value (NPV) for early focal enhancement to assess recurrent disease was 63 and 83%. The PPV and NPV for restricted diffusion were 83 and 80%. When both sequences were positive (4 of 21 patients), the PPV was 100% Table 6.

Table 6.

Late follow-up comparison of DCE and DWI in isolation and combination. DCE score: (1) Hypoenhancement, (2) isoenhancement and (3) hyperenhancement. DWI score: (1) No restricted diffusion and (2) restricted diffusion

Late follow-up	Late biopsy (median 53 months)
Combinations	Negative	Positive	All
DCE (3) and DWI (2) both negative	9 (82%)	2 (18%)	11 (100%)
Only DCE (3) positive	3 (75%)	1 (25%)	4 (100%)
Only DWI (2) positive	1 (50%)	1 (50%)	2 (100%)
DCE (3) and DWI (2) positive	0 (0%)	4 (100%)	4 (100%)
Total	13 (52%)	8 (48%)	21 (100%)

Paired PSA values were recorded in 17 of 21 men in the cohort Table 4. The mean PSA at median 4 months following treatment (new baseline or ‘nadir’) in the late follow-up cohort was 3.4 ng ml⁻¹ (range 0.6–7.5 ng ml⁻¹) and increased to 6.4 ng ml⁻¹ (1.0–16.4 ng ml⁻¹) in the follow-up period (median, 53 months). In males with positive biopsy, the mean PSA value changed from 2.12 ng ml⁻¹ (range 1.2–2.9 ng ml⁻¹) to 4.76 ng ml⁻¹ (range 1.6–9.6 ng ml⁻¹) while in males without csPCa, the mean PSA changed from 3.94 ng ml⁻¹ (range 0.6–7.5 ng ml⁻¹) to 6.99 ng ml⁻¹ (range 1–16.4 ng ml⁻¹) at late follow-up [p = 0.87] [Table 4.

Discussion

PSA plays a central role in the detection of recurrent tumour following whole gland therapy; however, it is less useful in patients treated with focal therapy wherein normal prostate tissue remains. ²² Recent studies have shown the superiority of mpMRI to PSA in detecting recurrent tumour in patients treated with FT. ¹⁸ However, using mpMRI to diagnose recurrent cancer specifically at the site of treatment in patients treated with focal therapy remains challenging as normal appearances post-treatment are less well defined. Although our cohort comprised of FLA, we expect the MRI findings to mimic that of other thermal ablation techniques such as HIFU.

Although 60% of the PCa sites harboured GG1 disease at baseline (mean CCL 0.85 mm), all sites were visible on MRI. Studies have shown MRI visibility in GG1 disease to be associated with upgrading at prostatectomy. ²³ Additionally, CCL of >6 mm has shown to be associated with adverse pathology on prostatectomy. ²⁴ Therefore, to evaluate the test characteristics of MRI following FT, GG1 ≤5 mm at follow-up biopsy was considered not clinically significant.

Our study shows that on T2W images ill-defined scarring with and without atrophy was most commonly encountered, 68 and 76% of cases at early and late follow-up, respectively. Well-defined scarring was far less common as was scarring with cystic change. Other studies reported an inability to visualise the ablation zone in certain cases on T2W images, ²⁵ however it was visible in all cases during our study. Interestingly, the different T2 appearances did not show any association with residual or recurrent disease at site of treatment.

Hötker et al documented oedema within the prostate at the ablation site visualised as intermediate T2 signal at a mean of 16- and 25-day post-treatment between two readers ultimately resolving by 252 and 514 days. ²⁶ This was not experienced in our study which is likely due to the fact that mpMRI review was performed no earlier than 3-month post-treatment.

Our data support other studies that have demonstrated a temporal change in the ablation zone on DCE. ²⁶ It is most commonly seen as a hypoenhancing defect initially, which becomes less conspicuous over time progressing to isoenhancement at late follow-up.

There was considerable overlap in the qualitative appearance of the prostate on T2W images. Most patients had some degree of atrophy with volume loss such that incorporating this into qualitative assessment is unlikely to be of use. As such, atrophy with volume loss could be considered an expected finding following focal therapy.

PSA is known to be particularly unreliable in the early follow-up period after FT. It is thought that following the initial decline / ‘nadir’ after FT, PSA may be a helpful marker in detecting recurrent disease at late follow-up. ²⁷ However, PSA change in the late follow-up period was not shown to be a helpful marker in our study. The small number of patients with recurrent disease and paired PSA values may have accounted for this.

Early mpMRI follow-up was performed at 4–8 months to avoid immediate post-ablation changes, which have been shown to mimic residual cancer. ^18,25 International multidisciplinary consensus statements suggest the initial MRI be performed within 6 months ²⁸ or upto 12 months following FT and periodically thereafter depending upon patient factors and resource availability ²⁹ which aligns with follow-up after FT at our institute. Our early follow-up results showed that mpMRI was able to detect recurrent tumour in only of one of four patients as focal early hyperenchancement. Other patients demonstrating focal hyperenhancement or restricted diffusion were negative at biopsy. All four patients with a positive biopsy did not demonstrate restricted diffusion. This may be related to small volume residual disease picked up on extended sampling (three to seven samples from ablation area including margins) in our protocol. It should be noted that one patient who showed both DCE hyperenhancement and DWI restricted diffusion but was negative at early biopsy went on to be diagnosed with GL7 disease at late follow-up 2.0 years post-treatment. This may have been from sampling/targeting error at time of the initial biopsy.

Late mpMRI was sensitive at detecting recurrent disease. Restricted diffusion was associated with recurrent disease in 83% of cases. Hyperenhancement was also associated with recurrent disease with biopsies positive in 63%. However, 25% of patients with positive biopsies did not show restricted diffusion or hyperenhancement. On the other hand, the PPV for both DCE hyperenhancement and restricted diffusion was 100%.

The study has several limitations. Sample size with recurrent PCa at both time points was small in the study and therefore these findings would require validation in a larger cohort. In earlier mpMRI scans, high B-values of over 1000 mm/s² were not obtained, and this may have contributed to lower sensitivity of the mpMRI in the early follow-up period. While the early follow-up MRI and biopsy were performed as per study protocol within a narrow time period of 4–8 months, late follow-up mpMRI and biopsy were performed on a clinical basis such that an element of bias may exist in this subset of patients. However, PSA change was not associated with biopsy results at either early or late follow-up, and hence it is unlikely to have affected the test characteristics of post-FT mpMRI.

Additionally, most of the early biopsies were performed with ‘cognitive fusion’ while the late biopsies were performed with MRI-TRUS fusion (Artemis 2.0, Eigen, California). This raises possibility of false negative results on early biopsy, although studies have shown similar PCa detection rates with different targeted biopsy techniques. ³⁰ Further, our results may apply only to targeted FT template for smaller cancers since tumours > 20 mm on MRI were not included in the trial.

Conclusion

Multiparametric MRI is able to document the morphological and temporal changes of the prostate following focal therapy. It has a limited ability to detect recurrent disease within the first 4–8 months following treatment. However, it is a sensitive marker at late follow-up where hyperenhancement and restricted diffusion are associated with recurrent disease.

Key points

• mpMRI demonstrates wide range of appearances of the prostate following focal therapy.

• At late follow-up after focal therapy, hyperenhancement and restricted diffusion were associated with recurrent disease.

• mpMRI is a sensitive marker for detecting recurrent prostate cancer after focal therapy.